Many dynamic systems can be characterized as complex since they have a nonlinear behaviour incorporating a stochastic uncertainty. It has been shown that one of the most appropriate methods for modelling of such systems is based on the application of Gaussian processes (GPs). The GP models provide a probabilistic non-parametric modelling approach for black-box identification of nonlinear stochastic systems. This book chapter reviews the methods for modelling and control of complex stochastic systems based on GP models. The GP-based modelling method is applied in a process engineering case study, which represents the dynamic modelling and control of a laboratory gas–liquid separator. GP models with different regressors and different covariance functions are obtained and evaluated. A selected GP model of the gas–liquid separator is further used to design an explicit stochastic model predictive controller to ensure the optimal control of the separator.
COBISS.SI-ID: 27718183
The article deals with a method for determining the correlation between the data on the spatial distribution of various atmospheric pollutants and data on the morbidity rate among school-aged children in the small Zasavje region above very complex terrain. We presented a modelling method for the assessment of air pollution, a data set on the morbidity rate among children and a method for determining suitable spatial units. We then searched for any causal relationship between the pollution and the morbidity rate within these units. The method takes into account the exact spatial diversity of the pollution and is intended to find unknown patterns (meaning that it does not draw conclusions on the basis of a preliminary definition of polluted and non-polluted spatial units) and then compare these patterns to the morbidity patterns.
COBISS.SI-ID: 38190125
One of the prerequisites for the proper determination of air pollution levels in a detailed temporal and spatial resolution is knowing the (spatially and temporally distributed) air pollution levels in the wider international area, or the so-called background air pollution level. In the scope of this project international modelling QualeAria system is used, which was developed by ARIANET in Milan, Italy. The research team has been collaborating with their research and development team in the field of air pollution modelling for two decades (both teams once operated as part of state institutions). The article presents a comprehensive validation of air pollution modelling results for Slovenia for a one-year period generated using the QualeAria system. The results are very encouraging and confirm that modelling results generated using the QualeAria operational system can be used as background pollution data for modelling systems for small regions in Slovenia.
COBISS.SI-ID: 28053799